A migration process is that a server device which performs a virtual machine moves this virtual machine to another server device, then the another server device performs this virtual machine (referring to patent documents 1).
[Patent documents 1] Japanese Laid-open Patent publication No. 2011-146002.
On the other hand, it is performed that a large number of the server device which use a business are distributed to two or more data centers in a company and a public institution, etc. This data center builds a virtual machine (VM) to one or more server device, and the virtual machine performs variety of information processing relating to the above described business. In addition, when the data center belongs to a service company, it is performed to provide users with the virtual machine which is built to the server device. This service is called to cloud service (IaaS: Infrastructure as a Service).
In the system which consisted of two or more of these data centers, it is effective that the server device allocated in the first data center migrates the virtual machine which is performed by the server device to another server device in the second data center which is isolated geographically from the first data center and operates independently from the first data center. For example, it is effective to perform migration for the purpose of the case where the first data center suffered a great deal of damage and fails for power, the case of reduction of cooling cost in the data center which used the seasonal difference of the area, and the case of operation cost reduction using low charge electric power at night, etc.
When the migration is performed between such data centers from which it is geographically separated, various restrictions arise. For example, it is assumed that the first data center suffers a great deal of damage, and is continued electric supply only for a short time. In this case, the first server device that is performing the virtual machine in the first data center is required to transmit this virtual machine to within a time which this electric supply is supplied.
However, for example, there is a case that the network between the first and second data centers is crowded, and a case the network between the first and second data centers may become narrow in bandwidth. In addition, there is a case that traffic may decrease due to an increase of the processing load in a certain network device (for example, switching device, router device) which is prepared between the first and second data center, although the bandwidth of the network is fully secured.
In such a case, because the network between the first and second data center serves as a bottleneck of data transmission, it becomes difficult that the first server device performs the migration at high speed. If it puts in another way, it will require time that the first server device in the first data center moves the virtual machine currently performed to the second server device in the second data center which is a destination, and the second server device builds the same virtual machine as this virtual machine.
Especially, it is difficult that the first server device in the first data center moves these virtual machines to the second server device in the second data center at high speed, when the first server device performs the virtual machines of the thousands and tens of thousands of units.